EMBER: Embedding-based multi-label prediction of phosphorylation events Developer: Kathryn E. Kirchoff from Gomez Lab (Computational Systems Biology & Network Medicine) at University of North Carolina—Chapel Hill.
A preprint mansucript describing EMBER and its application is available through bioRxiv: https://doi.org/10.1101/2020.02.04.934216
torch==1.3.1
numpy>=1.17.2
It is necessary to obtain the Siamese embedding because it is too large to be uploaded on github. Simply run the Siamese file. The model weights are saved. You may train your own network according to the code in the network.
Run the notebook to train EMBER network. Test data and training data utilized for the paper are provided.
You may test the network using our provided test set or use your own test set.
Please feel free to contact us if you need any help: kat@cs.unc.edu